Make it possible to interact with external data sources

[nils-braun]

Currently, all data frames need to be registered before using them in dask-sql.
However, it could also be interesting to have dataframes directly from S3 (or any other storage, such as hdfs), from the Hive metastore or by creating temporary views. In the background, one could create dask dataframes and use the normal registration process.

For this, we first need to come up with a good SQL syntax (which is supported by calcite) and/or an API for this.

[nils-braun]

It is now (after #55) possible to create new tables by reading in data, e.g. with

CREATE TABLE
        "nyc"
    WITH (
        format = 'csv',
        location = 'https://support.staffbase.com/hc/en-us/article_attachments/360009197031/username.csv',
        sep = ';'
    )

However, it is not clear so far if this suits all the needs (e.g. S3). Additionally, an interaction with hive would be interesting.

[raybellwaves]

BlazingDB/blazingsql#937 may also be relevant

[nils-braun]

Just a small update: interaction with hive is implemented as a experimetal feature in 0.2.0 (#63).
And as fsspec/adlfs#111 was implemented, it should also be possible to use az:// as input format (although untested)

[raybellwaves]

it should also be possible to use az:// as input format (although untested)

It'll be worth testing with a public repo

storage_options = {'account_name': 'azureopendatastorage'}
ddf = dd.read_parquet('az://nyctlc/green/puYear=2019/puMonth=*/*.parquet', storage_options=storage_options)

[raybellwaves]

see #84

[nils-braun]

Closing this issue now, as we have already integrations included and for the specific new types of integration, be better create a new issue (e.g. like #83)